Your AI model might disappear tomorrow: what now?
My
latest for LeadDev considers the effects of the Fable and Mythos quick shutdowns.
We all knew AI models could become outdated quickly. But the rapid shutdown of Fable and Mythos shows another risk: models can disappear altogether. That's why engineering teams need model-agnostic foundations.
Experts recommend considering open-source models and agents where possible, avoiding single-provider dependencies, and using abstraction layers between your workflows and underlying models.
"The AI market's velocity is also its vulnerability," says Robert Keus. "Model providers can pivot, deprecate, or disappear overnight."
My latest for LeadDev briefly reviews what happened last week and looks at what engineering teams can do next.
Frontier models have their place, but so does open infrastructure! That's a theme I'm excited to come back to soon with a dedicated feature, so stay tuned.











